While there has been great progress in understanding the risk architecture of autism, there are still unanswered questions about the nature of the genetic and non-genetic risk for autism. Many of these questions can be best addressed with a population-based epidemiological sample with detailed demographic and environmental information. To date, almost all studies on the etiology of autism relied on convenience samples, which are subject to biases in capturing genetic and, possibly even more so, environmental risk. Epidemiologically based samples provide a unique resource to identify genetic and non-genetic causes of autism, while allowing for a precise estimate of risk in the population attributed to each source of risk. Sweden benefits from a centralized medical system that has been the foundation of large-scale epidemiological studies in psychiatric disorders, particularly schizophrenia and bipolar disorder. In our opinion, the significance of this proposal lies in the value of a unique, population-based epidemiological sample, analyzed in such a way as to address several outstanding issues in autism. These include: 1) Better estimates of heritability and environment in autism; 2) assessing the rate of recurrent risk CNV in autism; 3) discovery of rare standing single nucleotide variation in autism; 4) dissection of mechanisms underlying the association of nongenetic findings with autism - and the discovery of novel environmental associations; and, 5) cross-disorder analyses to better understand shared liability to autism and schizophrenia. The aims are: 1) To ascertain and biobank at least 1300 cases with autistic disorder and 1000 additional controls, to develop an international resource for ASD, and to assess selected, putative risk factors; 2) To genotype all samples using high-density SNP arrays, including dense exome coverage, and sequence all trios using whole-exome approaches; and, 3) To use novel methods to assess the role of inherited and de novo variants in autism and to evaluate rare standing variation in autism, while integrating key environmental variables. In later years the relationship between autism risk and risk for schizophrenia will be assessed. The proposed research is innovative, in our opinion, because it ascertains autism samples in an epidemiologically-valid manner, targeting a genetically homogenous population, for which schizophrenia and bipolar samples have already been collected. The proposal is also innovative in the use of novel methods to estimate heritability and to identify rare, standing-variation conferring risk to autism, while providing an integrated model for genetics and environment in autism. Finally, the application is innovative, in our opinion, in that it provides the groundwork for understanding shared risk across autism and schizophrenia, making use of a homogenous group to have better power to identify shared risk. This new and substantively different approach to studying autism, compared to studies carried out in convenience samples, addresses many of the open questions in autism research and provides a path towards a better understanding of the risk factors for autism and ultimately to better interventions in autism.